Part 7 of 9 in After You Call: A Deep Dive on Philly 311

I have called Philadelphia's 311 system three times in three years about a leaning forty-foot tree on a derelict, abandoned house with a caved-in roof directly adjacent to mine — the kind of giant that would kill someone if it came down. Each ticket closed. The tree is still there. I am one entry in a much larger pattern: 1,058 Philadelphia addresses called 311 twenty or more times in the past two and a half years, and the system closed nearly every ticket. They are 0.5% of the city's addresses and they account for 6% of all 311 traffic.

I pulled every 311 service request from OpenDataPhilly between January 2024 and June 2026 — 553,308 geocoded tickets across 227,981 distinct addresses — and grouped them by address. The cache only goes back to 2024, so this is 2.5 years, not the full history. The pattern is sharp.

The Long Tail of Philadelphia

Most addresses that called 311 called once. A small minority called constantly.

Compare the two share columns. The top 0.5% of addresses by complaint volume — those 1,058 properties — account for 6.0% of all 311 complaints, twelve times their share of addresses. The bottom 51.7%, the single-complaint group, account for 21.3% of complaints, less than half their share. The same long-tail shape shows up in policing calls-for-service data, in restaurant health-inspection violations, in pothole work orders. The right tail: the 1,058 addresses where 311 has been called 20 or more times in 2.5 years.

The Inequality, in One Curve

The histogram shows the tail. The Lorenz curve measures it. The further the orange curve bows from the diagonal, the more concentrated the calls.

The Gini coefficient is 0.44 — in the range of U.S. income inequality (the income Gini is about 0.49). The top 1% of addresses, the 2,280 most-called properties in the city, generate 9.6% of all 311 calls. The top 5% generate 25%. The other 95% make the remaining 75% — the one-time-problem pattern.

The Gallery

The map below pins every Philadelphia address with 5 or more complaints — the full 25,500-property chronic set. Red dots are the worst (20+ calls), orange and amber are 10-19 and 7-9, yellow is 5 or 6. Click a dot for the address, neighborhood, total count, open count, and top service.

Three residential cases stand out. (Commercial and intersection addresses also show up in the chronic set — a different kind of chronic, and I come back to them.)

7901 Henry Avenue, Roxborough — 112 maintenance complaints, 66 still open

Between January 2024 and April 2026, 311 logged 112 maintenance complaints against 7901 Henry Avenue — an average of one every nine days, weekends included. Sixty-six are still open as of June 2026. A handful of other complaint types appear (abandoned-vehicle, street-defect, dangerous-building, traffic-signal), but the dominant signal is 112 maintenance tickets at one property over 835 days. That is not a service-mix artifact. That is a property the L&I system has cycled through 112 times in two years.

7901 Henry Avenue is not a single-family home. It is Henry on the Park, a Morgan Properties apartment complex in Roxborough, and the building carries a 1.5-out-of-5 average on Yelp across hundreds of reviews. Two representative reviews capture what chronic undermaintenance at a property like this looks like from the inside:

The shape of the 311 data — 112 tickets, mostly open, mostly the same type, two and a half years running — is the shape of those reviews. The 112 tickets are not a complaint from a homeowner. They are a record of what happens when a large rental property is chronically undermaintained, on-site staff are overworked, and the city's 311 line is the only escalation path the people inside the building have.

View the full 311 complaint log for 7901 Henry Avenue →

2114 South 8th Street, Whitman — 87 illegal-dumping reports

2114 South 8th Street logged 87 illegal-dumping reports over 709 days, every one closed. Illegal dumping is a chronic 311 problem citywide — 50,099 tickets in 2.5 years, most of them clustered on a small set of known hotspots — and 2114 S 8th is the worst single-service-dominant address for illegal dumping citywide — its 87 reports are all illegal dumping, the most single-service-dominant cluster in the city. The pattern is consistent with a vacant lot or problem property where dumping returns faster than the city can clear it. The 311 form default closure is that the cleanup was done. It does not record that the cleanup needs to be done again, and again.

View the full 311 complaint log for 2114 South 8th Street →

1415 Benner Street, Mayfair — 79 maintenance complaints

1415 Benner Street in Mayfair logged 79 maintenance complaints over 422 days, every one closed. The cluster sits on a residential block off Roosevelt Boulevard in Far Northeast Philadelphia. The complaint type is uniformly “Maintenance Complaint” — the same L&I bucket the previous article in this series identified as having the largest citywide income gap. The service mix mirrors 7901 Henry: a steady cadence of maintenance tickets the city records as closed but that a neighbor has to file again a few weeks later. The system is processing the work. The system is not solving the problem.

View the full 311 complaint log for 1415 Benner Street →

The Three Readings of the Data

I went looking for the corners 311 forgot. The data handed back something more complicated.

The first reading is what the gallery sub-articles above show: residential addresses where 311 has been called again and again for the same problem, and the system has cycled through the reports without breaking the cycle. 7901 Henry Avenue is the worst of these, but the gallery has 15.

The second reading is institutional address-of-record bleed. 1234 Market Street in Washington Square has 61 “KB Escalations” tickets from the city's KeySpot / Knowledge Bank program — 59 still open — plus 20 other service types. A Market Street address is the service-center of record on many of those tickets. That is the city's own data hygiene showing up at one address.

The third reading is intersection-as-address. The gallery includes Cresheim Valley Drive & Germantown Avenue in Chestnut Hill (76 street-defect reports, 28 still open), Aramingo Avenue & East Girard Avenue in Fishtown (70 homeless-encampment reports), and Fairmount Avenue & North 25th Street (56 traffic-signal emergencies). When a caller gives a corner, the geocoder places the call at the intersection — which aggregates every defect, signal, and encampment call from the surrounding blocks. The 76-ticket Cresheim Valley case is real. It is also a 76-ticket intersection, not a 76-ticket property.

The gallery is not a clean list of 15 forgotten corners. It is a mix: residential blocks with a real repeating problem; commercial and civic addresses 311 records at one canonical location; and intersections that aggregate calls from a wider area. Only the first should change policy. The other two are data hygiene and routing.

The Chronic Count Doesn't Skew Poor

The parent article found that lower-income zips wait longer for the same 311 service. I expected the chronic count to follow the same gradient — poor zips at the top, wealthy zips at the bottom. The data does not cooperate.

Of the 6,868 (address, service) pairs that hit the chronic threshold, 2,856 are in middle-income zips and 2,748 are in lower-income zips. The middle tier leads by 108 pairs — 1.6% of the chronic total, well within noise. Higher-income zips account for 872, and 392 are in zips that could not be mapped to a tier (mostly addresses whose zip code did not join cleanly to the ACS reference). The chronic count is roughly flat across lower and middle zips. It is in higher-income zips where the count drops off — about 3.1× less than lower.

Two readings. First: middle-income zips include a long stretch of West and Northwest Philadelphia with older housing stock that generates more chronic-311 signal per property. Second: lower- and middle-income zips together hold most of the city's housing, so the chronic set is naturally weighted there. Either way, the surprise holds: the chronic-same-service pattern is a citywide one, not a poverty-only one.

What the Status Notes Actually Say

The status_notes column is how the city communicates what happened on a 311 ticket. I read the notes for the gallery addresses. The most common closure reasons, in order, are:

The most striking finding is the leader. “Referred to another organization” is the single most common closure note in the gallery, on 81 of the 285 tickets that left a status note (28.4%). It is the system's polite phrase for “this is not our problem.” A maintenance ticket filed 30 times in two years can be closed every time by pointing the caller at a state agency, a non-profit, or another city department. The closure rate looks fine. The chronic case is now someone else's chronic case.

“Issue Resolved” is the second most common reason, on 79 tickets (27.7%), and it is the least informative. A 57-of-59-ticket graffiti cluster at 4800 Baltimore Avenue in Kingsessing closes every ticket the same way — and the same tag is removed over and over for two and a half years. The closure note does not say the underlying cause — a specific property, a specific surface that attracts the same tag — was addressed. The closure note just says it is clean right now.

“Information Provided” is the third most common, on 70 tickets (24.6%), and it dominates the abandoned-vehicle subset. It is the police department's default closure reason. It does not mean the vehicle was towed. It means a dispatcher told the caller what the city's abandoned-vehicle policy is. The 2540 South 17th Street address has 59 of these notes (out of 82 tickets), and 5 are still open. The pattern is: someone called, a dispatcher explained, the car stayed. The next caller files the same report.

One caveat: only 285 of the gallery's roughly 2,000 tickets have any status note at all. The 311 system writes a closure reason for less than 15% of closed cases. 1415 Benner Street, 2114 South 8th Street, and most of the residential illegal-dumping cases have zero notes — they're not in the chart, nothing to put in it. The bars above are a sample of how the system explains itself — when it bothers to. The dominant pattern in chronic cases is no explanation at all.

The pattern is not that the city ignores these addresses. The city's closure vocabulary is built for one-time work; long-cycle problems have no honest entry in it. A graffiti tag is a one-hour cleanup. A graffiti-tag problem is a property-management issue. The 311 data records the cleanup, not the problem.

Why the Cycle Continues

The 311 system is built around the one-call, one-resolution model. A caller files a report. A dispatcher assigns it to an agency. The agency closes it. The architecture is fine for a pothole or a street light, and it is the wrong architecture for a problem that keeps coming back. Chronic cases are filed as new tickets every time, with no link to the previous 49. The dispatcher who answers the 50th call at 7901 Henry Avenue cannot see that the dispatchers who answered the 1st, 25th, and 49th closed a case the same way. Each call starts a new case ID. Each case is judged on its own.

The parent article analyzed what happened with that same tree in Strawberry Mansion — a property the city knew about, had plans for, where the two bureaucracies that could act did not talk to each other. The chronic-address data is the same story, multiplied by however many of these properties exist in the city. The bureaucracy is not ignoring the problem. It is measuring itself on the wrong axis.

What Would Change It

Three things, in order of how much of the data they would touch.

First, a property-level flag in 311. When a (parcel ID, service type) combination has been reported N times in M months, the system should require a different closing reason than “Issue Resolved.” A property called 50 times for graffiti in two years is a property with a graffiti problem, not 50 separate one-time issues. The data should know that.

Second, intersection tickets routed to the nearest address. A 311 ticket filed at “X Ave and Y St” belongs to one corner of that intersection; the intersection itself is not a property. Routing to the nearest address would let the chronic-property analysis tell the difference between a property with 70 tickets and a corner with 70 tickets.

Third, a closure-reason vocabulary that admits when the problem is not solved. “Issue Resolved” on a graffiti tag that has been removed 57 times in two and a half years is a lie. “Issue Resolved” on a maintenance complaint the inspector could not locate is a lie. Adding a repeat-closure field would change how the city measures its work.

Notes, Sources, and Methodology

All data is from the OpenDataPhilly 311 Service Requests, covering 2024-01-01 through 2026-06-08. n = 553,308 geocoded requests across 227,981 distinct addresses. Chronic clusters are defined as 5+ reports of the same service at the same address, spanning at least 365 days; that produced 6,868 (address, service) pairs. The 2.5-year window reflects the period the public extract was available at the time of analysis, not a deliberate choice of window length.

One methodology choice that matters: a chronic cluster requires the same service, not just any repeat report. A property that has been called 50 times for graffiti and 5 times for illegal dumping counts as one chronic graffiti case and one non-chronic dumping case — not one 55-ticket case. The status_notes field is sparse — the 311 system writes a closure reason for only 13.4% of closed tickets citywide — which is why the closure-reason chart above is a sample of how the system explains itself, not a complete inventory of how the gallery was closed. Income tiers (lower / middle / higher) are the parent article's cutoffs, derived from ACS 2022 5-year zip-code median household income.